Magnetic suspension is useful for an implantable artificial heart because it offers high durability related to abolition of contacting mechanical parts. 1–5 We have been developing a magnetically suspended centrifugal blood pump since 1997. 4–6 A levitated impeller is suspended in the radial direction in our magnetically suspended centrifugal pump. This suspension makes the device thinner than other pumps because the electromagnetic system can be constructed around the impeller. Another issue in a magnetically suspended artificial heart is establishment of a pump structure to realize high performance and stable pumping. Efficiency, suspension stability, and pump performance are all strongly related to pump structure. In this study, we examined an impeller structure and a volute structure to achieve better suspension. A levitated motor was also redesigned to improve the performance of our device.
Materials and Methods
Magnetically Suspended Centrifugal Pump
Figure 1 shows a schematic view of the pump that is under development. The outer rotor mechanism has been adapted to permit levitation. The pump has a stator at the center and a levitated rotor at the circumference of the stator. The stator has two electromagnets for levitation control and rotation control, while the rotor has four permanent magnets on its inner surface. The outer diameter and height of the rotor-impeller are 63 and 34 mm, respectively. The rotor is levitated and rotated in the radial direction by the attractive force generated by the stator, and the axial movement of the rotor is passively restricted with magnetic force to reduce the number of control parameters. Two eddy current sensors were mounted to control the rotor position in the radial direction. The current supplied to the stator coils is controlled numerically with a digital signal processor. The levitated motor was driven as an AC servomotor with the rotational position signal. The gap between the rotor and the stator was set at 1 mm to prevent hemolysis. 7 An impeller is set on the levitated rotor in this device. We decided to add an open impeller to our pump by extending the vanes from the top surface of the outer rotor to the center of the pump. Before adopting the open impeller, the pump performance of the open impeller was compared with that of a semi-open impeller to test the performance of the open impeller. A double volute is a suitable way to passively compensate for fluid dynamic imbalance without extra power requirements, but the double volute structure increases pump production difficulties. Therefore, we compared the effect of a double volute on levitated impeller stability and power requirements with that of a single volute. The maximum rotating speed of the prototype pump was 2,000 rpm, determined by the performance of the levitation motor. We supposed that magnetic power was limited because magnetic flux saturation occurred in the stator of the prototype pump when the rotating speed exceeded 2,000 rpm. Stator geometry was redesigned to avoid this magnetic flux saturation by means of a 3-D finite element analysis of the magnetic circuit.
Open Impeller and Semi-Open Impeller
Figure 2 shows the tested impellers. Both impellers have six vanes and an outer diameter of 63 mm. The open impeller has no shroud between vanes, while the semi-open impeller has a back shroud between the vanes. There is a hole with a diameter of 10 mm at the center of the semi-open impeller for washout behind the impeller shroud. The ordinal magnetic coupling pump shown in Figure 3 was developed to examine the pump performance of both impeller pumps. The top of the pump casing is the same as that of the magnetic suspension pump. Sets of permanent magnets are placed on the impeller and drive sides. The permanent magnets on the drive side are rotated by a DC motor and the impeller, with another set of permanent magnets rotated along with it. A mock loop circuit filled with water was used to measure the pump performance.
Effectiveness of the Double Volute
The radial movement of the levitated impeller during pumping was evaluated with both the double and single volutes. The position of the levitated impeller during pumping at a rotating speed of 1,400 rpm was measured and calculated with eddy current sensors placed outside the impeller. Power consumption for the levitation control was observed simultaneously by measuring the power into the levitation control coils of the stator.
The stator geometry was redesigned to avoid this magnetic flux saturation. A 3-D finite element analysis of the magnetic circuit was performed to determine the location of the magnetic saturation and to estimate the effect of the geometric redesign. Commercial software (ANSYS, Canonsburg, PA) was used for the analysis. The stator and the rotor were modeled with finite elements, and magnetic flux was visualized. The previous stator and redesigned stator are shown in Figure 4. From the results of the 3-D finite element analysis, the thickness of the stator was increased to twice that of the previous one. The effectiveness of the redesign was measured by comparing the magnetic flux density and attraction produced by the stators. The magnetic flux density produced between the stator and the rotor was measured with a magnetic flux probe. Attractive force was directly measured with a strain gage sensor. Finally, the magnetically suspended centrifugal pump was constructed with the improved stator, the open impeller, and the double volute. Its pump performance and efficiency were evaluated with a mock loop circuit filled with water.
Pump Performance with the Open and Semi-Open Impeller Pumps
Head pressure–flow rate (HQ) characteristics of the magnetic coupling pumps with an open or semi-open impeller are shown in Figure 5. The open impeller showed better performance, although the difference between the two was small.
Effectiveness of the Double Volute
The trajectory of the center of the levitated impeller during pumping is shown in Figure 6. The levitated impeller was stabilized within 0.05 mm with the double volute; the control range of the impeller with the single volute was within 0.07 mm. Power consumption for the levitation control is shown Figure 7. At a rotating speed of 1,500 rpm, it was 0.7 W for the double volute pump; for the single volute, it was 1.3 W.
Motor Performance and Pump Performance
Figure 8 shows the visualization results of the magnetic flux in the stators analyzed using a 3-D finite element method. Brighter parts in Figure 8 indicate the location of magnetic saturation. The previous stator exhibited saturation at the spokes of poles. The poles have a bulge at the end to distribute the magnetic field effectively and narrow spokes to wind the coils. Therefore, the cross-sectional area in the poles decreases from the bulge to the spoke. The magnetic flux concentrated from the large bulge area to the narrow spoke area, with the magnetic flux saturation occurring at the boundary of the bulge. We increased the thickness of the spoke part by two-fold in the redesigned stator, and saturation of the magnetic flux was avoided (Figure 8B). The improvement in the magnetic flux density is shown in Figure 9. Flux density of the new stator increased and the linearity of the flux change against the excitation current was also improved. Figure 10 is a comparison of the attraction of the stators. Attractive force was also increased in the redesigned stator. HQ characteristics and efficiency of the levitated pump with the redesigned stator is shown in Figure 11. Maximum rotating speed increased to 2,200 rpm. Maximum head pressure and flow rate were also increased to 250 mm Hg and 9 L/min, respectively. Efficiency with a head pressure of 100 mm Hg and a flow rate of 5 L/min was 11%.
The back shroud of the centrifugal blood pump allows blood to stagnate behind the impeller and is a candidate area for thrombosis formation. Establishment of better washout conditions behind the impeller is an essential issue in developing a successful centrifugal blood pump. The blood stagnation area at the back surface of the impeller shroud could be minimized with an open impeller structure, but it is possible that the regurgitation between the vanes would decrease pump performance. Still, an open impeller limits the thrust force produced in the centrifugal pump. Minimizing the thrust force is important for stable suspension of the impeller because the axial movement of the levitated impeller is restricted only by passive stability in our pump. Thus, it is possible that the levitated impeller would move in the axial direction with large thrust forces. We expected the semi-open impeller to perform better than the open impeller. The experimental results, however, indicate better performance for the open impeller, the reason being that the open impeller has lower friction loss than the semi-closed impeller, while mechanical loss from regurgitation without the back shroud did not reduce pump performance. These results confirm the suitability of the open impeller.
The radial movement of the rotor is regulated actively in our pump. This means that the pump can compensate for the imbalance of radial force in a centrifugal pump. Using a double volute is another common mechanical way to passively compensate for fluid dynamic imbalance without extra power requirements. Although a single volute pump is simpler and easier to produce, we decided to evaluate the effect of a double volute on levitated impeller stability and power requirements. The double volute pump with the magnetic coupling produced better stability in the levitated impeller and slightly higher efficiency, but the difference was small. In the magnetic suspension experiment, movement of the levitated impeller with a single volute pump was controlled within 0.07 mm, and power consumption was only 1.4 W. The input power difference between the double and single volute pumps was only 0.7 W. This means that the simpler single volute could be used in the magnetically suspended centrifugal blood pump.
The reason the magnetic saturation was located at the spokes of the poles in the previous stator is as follows. The poles have a bulge at the end to distribute the magnetic field effectively and narrow spokes to wind the coils. Therefore, the cross-sectional area in the poles decreases from the bulge to the spoke. The magnetic flux concentrated from the large bulge area to the narrow spoke area, with magnetic flux saturation occurring at the boundary of the bulge. Changing the thickness of the stator spoke enlarged the cross-sectional area of the spokes, prevented saturation of the magnetic flux, and improved the motor’s performance. This improvement was evident in the flux change linearity and attraction. The redesigned pump indicated HQ characteristics similar to those of the previous pump. However, the maximum rotating speed of the levitated pump was increased from 2,000 rpm in the previous pump to 2,200 rpm. The efficiency of the new pump was also similar to that of the previous pump. Although extra power was required for levitation control, the efficiency of this levitated pump was still high enough for implantation. Pump performance will decrease to about 90% of the in vitro test results with water when the pump is used as a blood pump because of fluid viscosity differences. The estimated maximum flow rate using the current pump as a blood pump is 8 L/min. We believe that a left ventricular assist device suitable for implantation should have the following characteristics: the maximum head pressure should be 250 mm Hg, the maximum flow rate should be at least 8 L/min, and power consumption should be less than 15 W with a flow rate of 5 L/min against a pressure head of 100 mm Hg. Because our redesigned pump meets these criteria, we believe that the performance of our device is sufficient for use as a left ventricular assist devise.
The next step in this study is confirmation of the antithrombogenic and hemolytic properties. In this pump, the blood stagnation area at the impeller was minimized with an open impeller structure, but there are still a few problems concerning the area at the bottom of the rotor. We believe the shear stress at the rotor bottom is large enough to prevent thrombosis, and the gap between the impeller and the casing was designed based on the results of hemolysis studies. 7 It is assumed that the current pump design will be good at preventing hemolysis. Blood compatibility of the developed pump will be examined in animal experiments in the near future.
Pump performance of a centrifugal pump using an open impeller was compared with one using a semi-open impeller. The open impeller produced better pump performance than the semi-open impeller, and the effectiveness of a double volute was also confirmed. However, the single volute pump also had good stability in the levitated impeller and low power consumption requirements for levitation control. The stator of the levitated motor was redesigned to avoid magnetic saturation and improve motor performance. The changes were effective, and performance of the new pump with the redesigned stator was improved. The developed magnetically suspended centrifugal blood pump is a candidate for an implantable left ventricular assist device.
This work was supported financially, in part, by a Grant-in-Aid for Scientific Research (B) from Japan Society for the Promotion of Science (no. 12555062).
1. Nojiri C, Kijima T, Maekawa J, et al: More than 1 year continuous operation of a centrifugal pump with a magnetically suspended impeller. ASAIO J 43: M548–552, 1997.
2. Allaire P, Hilton E, Baloh M, et al: Performance of a continuous flow ventricular assist device: magnetic bearing design, construction, and testing. Artif Organs 22: 475–480, 1998.
3. Schoeb R, Barletta N, Fleischli A, et al: A bearingless motor for a left ventricular assist device (LVAD), in Proceedings of 7th International Symposium on Magnetic Bearings
, Zurich, Switzerland, August 23–25, 2000, pp. 383–388.
4. Masuzawa T, Kita T, Matsuda K, Okada Y: Magnetically suspended rotary blood pump with radial type combined motor-bearing. Artif Organs 24: 468–474, 2000.
5. Masuzawa T, Toshiyuki Kita, Okada Y: An ultradurable and compact rotary blood pump with a magnetically suspended impeller in the radial direction. Artif Organs 25: 395–399, 2001.
6. Okada Y, Dejima K, Ohishi T: Analysis and comparison of PM synchronous motor and induction motor type magnetic bearings. IEEE Trans Ind Appl 31: 1047–1053, 1995.
7. Masuzawa T, Tsukiya T, Endo S, et al: Development of design methods for a centrifugal blood pump with a fluid dynamic approach: Results in hemolysis tests. Artif Organs 23: 757–761, 1999.